Ashish Gupta

Ashish Gupta

Computer scientist, founder, and engineering leader.

Ashish Gupta

I like building intelligent systems for hard problems. My career has sat where algorithms, data, and real decisions meet: building hedging and risk systems for equity options on Wall Street, leading machine learning for Google's ad auctions across billions of clicks, and co-founding companies in data and healthcare.

Today I'm co-founder and Head of Agentic Engagement at Arbiter, where we're building a care orchestration platform that unites healthcare.

Outside of work, I run a quantitative investing initiative: making the risk thinking I learned on Wall Street accessible to everyday investors.

Career

  1. Arbiter

    Co-Founder, Head of Agentic Engagement 2026 – present

    A care orchestration platform that unites healthcare.

  2. Kain Capital

    Chief Technology Officer 2023 – 2025

    A healthcare-focused private equity firm that partners with founder-led, mid-sized healthcare companies and transforms them through information technology. Led technology and strategy initiatives across the firm and its portfolio. New York.

  3. Polymer Search

    Co-Founder; CEO, then CTO 2019 – 2023

    The first no-code tool to make structured data accessible to anyone.

  4. Google

    Technical Lead, AdWords Optimization & Machine Learning 2012 – 2018

    Led conversion optimization and machine learning for Google Search Ads: models to measure, predict, and optimize auction outcomes, and new metrics for advertiser value, measured and optimized over billions of clicks.

  5. D.E. Shaw & Co.

    Quantitative Developer, Options Trading 2008 – 2012

    Built electronic trading systems for equity options: portfolio hedging, risk management, trade execution, forecasting, and real-time analytics.

  6. Google

    Engineering Intern 2006

    AdSense: a new ad delivery algorithm and a system for boosting ad relevance and click rates.

  7. Intel Labs

    Research Intern 2005 and 2006

    Pittsburgh (with CMU) and Cambridge, UK: distributed algorithms for modular robotics, and high-performance packet classification.

  8. IBM Research

    Research Intern Summer 2002

    Tools for building large-scale distributed e-government applications. New Delhi.

Education & research

  1. Northwestern University

    PhD, Computer Science 2002 – 2008

    Thesis: Black Box Methods for Inferring Parallel Applications' Properties in Virtual Environments, advised by Prof. Peter Dinda.

  2. IIT Delhi

    B.Tech, Computer Science 1998 – 2002

PhD, publications, and patent →

Quantitative investing initiative

A long-running project: making the quantitative thinking I learned at D.E. Shaw, such as volatility, drawdowns, correlation, and allocation, accessible to everyday investors, under QuantCommons.

Building for fun

I started writing software, and selling some of it, at twelve, including programs for teaching astronomy and physics. Later came programming contests, including the ACM ICPC World Finals in Vancouver in 2001, and games with AI opponents that were very hard to beat.

Projects, old and new →